Fibroblast Growth Factor-2 Represses Platelet-derived Growth Factor Receptor-α (PDGFR-α) Transcription via ERK1/2-dependent Sp1 Phosphorylation and an Atypical cis-Acting Element in the Proximal PDGFR-α Promoter*

Platelet-derived growth factor (PDGF) is a potent mitogen and chemoattractant for vascular smooth muscle cells (SMCs) whose biological activity is mediated via its high affinity interaction with specific cell surface receptors. The molecular mechanisms governing the expression of PDGF receptor-α (PDGFR-α) are poorly understood. Here we demonstrate that PDGFR-α protein and transcriptional regulation in SMCs is under the positive regulatory influence of the zinc finger nuclear protein, Sp1. Electrophoretic mobility shift, competition, and supershift analysis revealed the existence of an atypical G-rich Sp1-binding element located in the PDGFR-α promoter –61 to –52 bp upstream of the transcriptional start site. Mutation of this sequence ablated endogenous Sp1 binding and activation of the PDGFR-α promoter. PDGFR-α transcription, mRNA, and protein expression were repressed in SMCs exposed to fibroblast growth factor-2 (FGF-2). This inhibition was rescued by the blockade of extracellular signal-regulated kinase-1/2 (ERK1/2). FGF-2 repression of PDGFR-α transcription was abrogated upon mutation of this Sp1-response element. FGF-2 stimulated Sp1 phosphorylation in an ERK1/2- but not p38-dependent manner, the growth factor enhancing Sp1 interaction with the PDGFR-α promoter. Mutation of residues Thr453 and Thr739 in Sp1 (amino acids phosphorylated by ERK) blocked FGF-2 repression of PDGFR-α transcription. These findings, taken together, demonstrate that FGF-2 stimulates ERK1/2-dependent Sp1 phosphorylation, thereby repressing PDGFR-α transcription via the –61/–52 element in the PDGFR-α promoter. Phosphorylation triggered by FGF-2 switches Sp1 from an activator to a repressor of PDGFR-α transcription, a finding previously unreported in any Sp1-dependent gene.

transcription factor Sp1 regulates transcription of the PDGF-A and PDGF-B genes in SMCs and in other vascular cell types (20 -22). Here we show that PDGFR-␣ expression in SMCs is under the positive regulatory influence of Sp1, via an atypical recognition element located Ϫ61/Ϫ52 bp upstream of the transcriptional start site. This site mediates FGF-2 repression of PDGFR-␣ expression in an ERK1/2-dependent manner. Phosphorylation of Sp1 upon exposure to FGF-2 switches Sp1 from an activator to a repressor of PDGFR-␣ transcription.
Transient Transfections-For reporter gene analysis, WKY12-22 cells were transfected with 10 g of pLuc-a2, pLuc-a2.Sp1m3, or CMV-FIG. 1. Sp1 stimulates PDGFR-␣ protein and promoter-dependent expression. A, Western blot analysis for PDGFR-␣ in SMCs 24 h after transfection with CMV-Sp1 or pcDNA3. The non-specific (ns) band of lower molecular mass demonstrates unbiased loading. As shown in B, CMV-Sp1 increases PDGFR-␣ promoter activity in a dose-dependent manner. The backbone control, pcDNA3, has no effect. Firefly luciferase activity was normalized to Renilla. Error bars represent S.E. The data are representative of two or more independent determinations.
FIG. 2. Nucleotide sequence of the PDGFR-␣ proximal promoter. The transcriptional (transc) start site (ATG) (23) is indicated. The novel Sp1-binding element that is the subject of this study (Ϫ61/ Ϫ52) is indicated. Consensus elements for Sp1 do not appear in this region of the promoter. The C/EBP element and putative site for AP-1 are indicated. The sequence was obtained from EMBL RN13172 Rattus norvegicus.
FIG. 3. Atypical Sp1-binding element in the PDGFR-␣ promoter interacts with endogenous Sp1. This figure shows an electrophoretic mobility shift assay using 32 P-labeled oligonucleotide spanning the Ϫ80/Ϫ33 region ( 32 P-Mutant PDGFR-␣ Oligo Ϫ80/Ϫ33) of the PDGFR-␣ promoter or the 32 P-labeled mutant (G 10 3 T 10 ) oligonucleotide and SMC nuclear extracts. Supershift analysis with Sp1 antibodies abrogates complexes C 1 and C 3 and decreases intensity of C 2 and C 4 , whereas Ets-1 antibodies have no effect. UL, unlabeled oligonucleotide. Sp1.mThr 453 /mThr 739 . Cells were also transfected with 0.5 g of pRL-TK to correct for transfection efficiency. Firefly luciferase activity was normalized to Renilla. Transient transfections were performed using FuGENE6 (Roche Applied Science) where 3 l of FuGENE6/g of transfected DNA was added, and the transfection mix was made up to 1 ml with serum-free medium. After incubation at 22°C for 10 min, the DNA/FuGENE6 mixture was added to cells containing 10 ml of complete medium. Twenty-four h after transfection, cell lysates were prepared for assessment of luciferase activity where the dual luciferase assay reporter system (Promega DLR™) was performed on a manual luminometer (model TD-20/20 Turner Designs, Quantum Science).
Electrophoretic Mobility Shift Assay-Binding reactions for gel shift assays were performed in 20 l of 10 mM Tris-HCl, 50 mM MgCl 2 , 1 mM EDTA, 1 mM DTT, 5% glycerol, 1 mM PMSF, 1 g of salmon sperm DNA (Sigma), 32 P-labeled oligonucleotide probe (150,000 cpm), and 3 g of nuclear extract (determined by Pierce protein assay). The reaction was incubated for 35 min at 22°C. In supershift studies, 2 l of the appropriate affinity-purified anti-peptide polyclonal antibody (Santa Cruz Biotechnology) was incubated with the binding mix for 10 min before the addition of the probe. Bound complexes were separated from free probe by loading samples onto a 6% non-denaturing polyacrylamide gel and electrophoresing at 120 V for 2.5 h. The gels were vacuum-dried at 80°C and subjected to autoradiography overnight at Ϫ80°C. ؊52 abrogates Sp1-inducible PDGFR-␣ promoter activity. Transient transfection analysis in SMCs using pLuc-a2 or pLuc-a2.Sp1m3 together with CMV-Sp1 or pcDNA3 demonstrates that Sp1-inducible PDGFR-␣ expression is no longer observed in pLuc-a2.Sp1m3. Firefly luciferase activity was normalized to Renilla. Error bars represent S.E. The data are representative of two or more independent determinations.

Sp1 Positively Regulates PDGFR-␣ Transcription and Protein Expression-Our
FIG. 5. FGF-2 inhibits PDGFR-␣ expression. As shown in A, FGF-2 inhibits PDGFR-␣ promoter-dependent activity. SMCs were transfected with pLuc-a2 and treated with the indicated concentration of FGF-2 for 24 h prior to the assessment of luciferase activity. As shown in B, FGF-2 inhibits PDGFR-␣ mRNA expression. Reverse-transcriptase-PCR of PDGFR-␣ using SMCs was performed following FGF-2 treatment. GAPDH expression demonstrates unbiased loading. Firefly luciferase activity was normalized to Renilla. Error bars represent S.E. The data are representative of two or more independent determinations. ments in the proximal promoter regions of these genes. Unlike the PDGF-A and -B promoters, however, consensus elements for Sp1 do not appear in the proximal PDGFR-␣ promoter. Whether Sp1 controls PDGFR-␣ expression has not been investigated in any cell type. We assessed levels of PDGFR-␣ protein in vascular SMCs 24 h after transfection with the CMV-based Sp1 expression vector, CMV-Sp1. Western immunoblot analysis revealed that Sp1 overexpression produced a discreet band of molecular mass 170 kDa (Fig. 1A), which was barely apparent in cells transfected with the backbone vector, pcDNA3 (Fig.  1A). To confirm these observations at the level of transcription, we co-transfected SMCs with CMV-Sp1 and pLuc-a2, a Firefly luciferase-based reporter construct driven by 1.3 kb of PDGFR-␣ promoter (23). The cells were also transfected with the Renilla luciferase-based construct to correct for transfection efficiency. Normalized luciferase activity 24 h after transfection revealed dose-dependent induction of PDGFR-␣ transcription by Sp1 (Fig. 1B).
An Atypical Sp1-binding Motif in the PDGFR␣ Promoter Serves as a Functional Sp1-response Element-The proximal region of the PDGFR-␣ promoter does not contain a consensus Sp1-binding motif (5Ј-GGGCGG-3Ј). However, a G-box comprising 10 consecutive guanines was present at position Ϫ61 G 10 Ϫ52 , relative to the transcriptional start site (Fig. 2). To determine whether this site could support an interaction with Sp1, we performed an electrophoretic mobility shift assay using SMC nuclear extracts and a 32 P-labeled double-stranded oligonucleotide ( 32 P-PDGFR-␣ oligonucleotide Ϫ80/Ϫ33) whose sequence spans bp Ϫ80/Ϫ33 in the PDGFR-␣ promoter. This produced a number of nucleoprotein complexes (C 1 -C 4 ) whose specificity was demonstrated by their abrogation in the presence of a 150-fold molar excess of unlabeled oligonucleotide (Fig. 3, UL). The same -fold excess of an irrelevant oligonucleotide containing the NF-B-binding site failed to change the profile of these complexes (Fig. 3). Transverse mutation of the 5Ј-G 10 -3Ј ( Ϫ61 G 10 Ϫ52 to Ϫ61 T 10 Ϫ52 ) motif in the oligonucleotide abrogated complex formation (Fig. 3, 32 P-Mutant PDGFR-␣ Oligo Ϫ80/Ϫ33). Incubation of extracts with antibodies to Sp1 eliminated complexes C 1 and C 3 and decreased the intensity of complexes C 2 and C 4 (Fig. 3), whereas antibodies to Ets-1 had no effect (Fig. 3). Thus, endogenous Sp1 protein interacts with this atypical Sp1 element in the PDGFR-␣ promoter in a sequence-specific manner.
To establish the functional significance of this atypical Sp1binding element in the context of 1.3 kb of the PDGFR-␣ promoter, the same transverse mutation was introduced into pLuc-a2 (pLuc-a2.Sp1m3). Co-transfection analysis with CMV-Sp1 or pcDNA3 in SMCs together with mutant pLuc-a2.Sp1m3 completely abrogated Sp1 induction of the PDGFR-␣ promoter (Fig. 4). These findings thus demonstrate the existence of a novel Sp1-response element in the proximal region of the PDGFR-␣ promoter.
FGF-2 Represses PDGFR-␣ Transcription, mRNA, and Protein Expression-PDGFR-␣ levels progressively decrease in the FIG. 6. FGF-2-repression of PDGFR-␣ is ERK1/2-dependent. As shown in the left panel, FGF-2 suppression of PDGFR-␣ is rescued following treatment with PD98059. Western blot analysis using total cell extracts of SMCs exposed to FGF-2 for 8 or 24 h is shown. Cells were incubated with PD98059 (10 M) or SB202190 (500 nM) for 1 h prior to the addition of FGF-2. As shown in the right panel, the ERK1/2 inhibitor rescues the PDGFR-␣ promoter from repression by FGF-2. Luciferase activity was measured following 24 h of FGF-2/PD98059 treatment. Firefly luciferase activity was normalized to Renilla. Error bars represent S.E. The data are representative of two or more independent determinations. FIG. 7. The ؊61 G 10 ؊52 element in proximal PDGFR-␣ promoter is critical for FGF-2 repression of PDGFR-␣ transcription. Transient transfection analysis in SMCs using pLuc-a2 and pLuc-a2.Sp1m3 together with FGF-2 was performed. Firefly luciferase activity was normalized to Renilla. Error bars represent S.E. The data are representative of two or more independent determinations. growing atheroma as levels of FGF-2 increase (12,13). We hypothesized that FGF-2 may negatively influence PDGFR-␣ expression. Whether a growth factor can repress the expression of the receptor of another has not been demonstrated previously. Luciferase activity in SMCs transfected with pLuc-a2 was reduced by FGF-2 in a dose-dependent manner within 24 h (Fig. 5A). Semiquantitative reverse-transcriptase-PCR analysis confirmed these data. FGF-2 repressed endogenous PDGFR-␣ mRNA expression 24 h after exposure of the cells to the growth factor (Fig. 5B, left panel). Corresponding GAPDH transcript levels demonstrated unbiased sample loading (Fig.  5B, right panel). Western blot analysis further revealed FGF-2 suppression of PDGFR-␣ protein expression, after 8 and 24 h of incubation with FGF-2 (Fig. 6, left panel). These findings indicate that FGF-2 inhibits PDGFR-␣ gene expression at the level of mRNA and protein.

FGF-2 Repression of PDGFR-␣ Gene Expression Is
Dependent on ERK1/2 but Not p38 MAP Kinase-We and others have demonstrated that a major signaling pathway used by FGF-2 in SMCs is the extracellular signal-regulated kinase-1/2 (ERK1/2) cascade (24 -27). We hypothesized that FGF-2 repression of PDGFR-␣ involves the ERK1/2 pathway. Western blot analysis revealed that FGF-2 repression of PDGFR-␣ protein expression was rescued by the MEK/ERK inhibitor PD98059 (10 M) at 8 and 24 h (Fig. 6, left panel). In contrast, the p38 kinase inhibitor SB202190 (500 nM) failed to modulate FGF-2 downregulation of PDGFR-␣ expression (Fig. 6, left panel). Reversibility of FGF-2 inhibition by the ERK1/2 inhibitor, but not the p38 axis represents the ratio of upper to lower Sp1 band following Western blot analysis and densitometry. As shown in C, FGF-2 stimulates the interaction of Sp1 with PDGFR-␣ promoter. An electrophoretic mobility shift assay using nuclear extracts of SMCs treated with FGF-2 demonstrates the formation of at least one nucleoprotein complex that is supershifted with Sp1 antibodies, whereas Smad2 antibodies have no effect. 32 P-Mutant PDGFR-␣ Oligo Ϫ80/Ϫ33, 32 P-labeled oligonucleotide spanning the Ϫ80/Ϫ33 region.
inhibitor, was also demonstrated at the level of transcription in transfection analysis using pLuc-a2 (Fig. 6, right panel).

FGF-2 Repression of PDGFR-␣ Transcription via Sp1
Repression Element-Previous studies have shown that activation of ERK1/2 can phosphorylate and alter the interaction of Sp1 with the promoters of responsive genes (28). For example, ERK1/2-phosphorylation of Sp1 increases its interaction with the gastrin gene promoter and augments epidermal growth factor-induced gastrin gene expression (29). Whether ERK1/2inducible Sp1 phosphorylation mediates repression of gene expression is presently not known. Since FGF-2 activates ERK1/2 (27) and PD98059 reverses FGF-2 repression of the PDGFR-␣ promoter (Fig. 6, right panel), we hypothesized that FGF-2 inhibition of PDGFR-␣ is mediated via modification of Sp1. To begin to address this possibility, we exposed SMCs transfected with pLuc-a2.Sp1m3 to FGF-2 and assessed luciferase activity after 24 h (Fig. 7). FGF-2 down-regulation of PDGFR-␣ promoter activity was virtually abrogated by the mutation in the Sp1binding site (Fig. 7). This novel Sp1-response element therefore mediates FGF-2 repression of the PDGFR-␣ promoter. We next investigated the effect of FGF-2 at the level of Sp1 protein.

FGF-2 Stimulates Sp1 Phosphorylation in an ERK1/2-dependent Manner Increasing Sp1
Interaction with the PDGFR-␣ Promoter-Phosphorylation of Sp1 increases its molecular mass; the hyper-and hypo-phosphorylated forms can be readily resolved by polyacrylamide gel electrophoresis (30,31). Western blot analysis for Sp1 using nuclear extracts from SMCs treated with FGF-2 for 8 and 24 h demonstrated increased levels of Sp1 phosphorylation (Fig. 8A). Incubation of extracts from FGF-2-treated cells with calf intestinal phosphatase (CIP) eliminated the hyperphosphorylated form of this transcription factor, whereas incubation with bovine serum albumin (BSA) had no effect (Fig. 8A). These findings indicate, for the first time, that FGF-2 stimulates Sp1 phosphorylation.
The transcriptional mechanisms regulating PDGFR-␣ expression are poorly understood. In this study, we demonstrate that PDGFR-␣ transcription and protein expression are activated by native Sp1, which binds to an atypical G-rich binding element located at Ϫ61/Ϫ52 bp in the PDGFR-␣ promoter. Interestingly, this element also mediates FGF-2 inhibition of PDGFR-␣ activity, which is reversed by mutation of the cisacting element or blockade of ERK1/2 but not p38. Mutation of residues Thr 453 and Thr 739 in Sp1 that are phosphorylated by ERK1/2 abrogates FGF-2 repression of PDGFR-␣ transcription. Although native Sp1 activates PDGFR-␣, Sp1 phosphorylation by FGF-2 increases its interaction with the PDGFR-␣ promoter, reducing its activity via this cis-acting element. Thus, Thr 453 /Thr 739 phosphorylation switches Sp1 from an activator to a repressor of PDGFR-␣ transcription. This is the first report of Sp1 phosphorylation by FGF-2. It also demonstrates the capacity of a growth factor to regulate the expression of the receptor of another, as has been observed for the insulin-like growth factor-I receptor with PDGF (33) and FGF-2 (34). It is thus tempting to speculate that the inverse relationship between the expression of FGF-2 and PDGFR-␣ in the complex milieu of a developing atheroma may be mediated by the phosphorylation status of Sp1 and this novel recognition element in the PDGFR-␣ promoter.